305-300 Practice Questions

The docker stats command provides a live stream of resource usage statistics for running containers. This includes CPU usage, memory usage, network I/O, and crucially for this scenario, block I/O (disk reads and writes). This allows an administrator to quickly identify if a container is performing an excessive amount of disk activity. docker top shows processes inside a container, docker events shows daemon events, and docker system df shows overall disk usage by Docker objects.

False. The statement reverses the concepts. A Para-Virtualized (PV) guest has a modified kernel that is aware it is being virtualized, allowing it to communicate directly with the hypervisor for better performance in some cases. A Hardware-assisted Virtual Machine (HVM) guest uses CPU virtualization extensions (like Intel VT-x or AMD-V) to run a completely unmodified operating system, as the hardware provides the necessary abstraction.

To enable the QEMU Guest Agent, you need a three-part configuration. First, -device virtio-serial-pci adds the virtio serial device to the VM. Second, -chardev socket,... creates a character device on the host, in this case a Unix socket. Third, -device virtserialport,... connects the virtio serial device inside the VM to the character device on the host. This combination establishes the communication channel for the guest agent. The other options are incomplete or use incorrect device names.

Docker uses kernel cgroups to enforce resource limits. The --memory (or -m) flag sets a hard limit on the amount of RAM the container can use. To prevent the container from using swap space when it hits its memory limit (which can lead to performance degradation), you should set --memory-swap to the same value as --memory. This effectively disables swapping for the container. ulimit sets limits per-process, not for the entire container cgroup, and --shm-size relates to shared memory, not main RAM usage.

The correct cloud-config syntax, which starts with the #cloud-config shebang, uses declarative, top-level keys like users, packages, and write_files. This option correctly uses a list for the users module to define the new user and their sudo privileges, a list for the packages module to specify packages for installation, and the write_files module to create a file with specified content.

The virsh numatune command is specifically designed to display and modify NUMA parameters for a running virtual machine. This command provides details on memory node binding and vCPU to physical CPU pinning, which is essential for diagnosing and resolving NUMA-related performance issues. vcpuinfo shows vCPU state but not pinning, nodeinfo shows host NUMA details but not the VM's configuration, and memtune controls overall memory limits, not NUMA placement.

Multi-stage builds allow you to use one stage with a full SDK to build the application (e.g., golang:1.19) and a subsequent stage to copy only the compiled binary into a minimal base image (e.g., scratch or alpine). This ensures that build-time dependencies, compilers, and source code are not present in the final image, dramatically reducing its size and attack surface. It does not automatically handle image signing or affect the build cache in the described manner.

The correct command is qemu-img convert -f vmdk -O qcow2 -c -o preallocation=metadata source.vmdk target.qcow2. The -f flag specifies the source format, -O specifies the output format. The -c flag is the shorthand for enabling qcow2 compression. The -o flag is used for format-specific options, and preallocation=metadata tells qemu-img to create a non-sparse image where only the metadata is preallocated, which can improve performance by avoiding metadata fragmentation.

LXD clustering relies on communication over TCP port 8443 for its API and internal raft consensus protocol. While ICMP (ping) may be allowed, a common security practice is to block most incoming ports by default. A network timeout during the join process, despite successful ping tests, strongly indicates that a firewall is blocking the specific port LXD requires for cluster communication. Storage pool names being different or NTP issues would typically result in different error messages after the initial connection is established.

False. By default, the vagrant-libvirt provider will attempt to connect the VM to libvirt's 'default' virtual network, which is typically a NAT-based network. This allows the VM to access the external network and provides a DHCP-assigned IP address. While defining a private network (config.vm.network "private_network") is a common practice for stable host-to-guest communication, it is not strictly required for the VM to get network connectivity upon creation.